{
  "id": 243294,
  "title": "So what was Private and Public LB splitting logic?",
  "url": "/competitions/birdclef-2021/discussion/243294",
  "author_name": "",
  "post_date": "2021-06-02T00:26:27.910610Z",
  "votes": 4,
  "comment_count": 2,
  "views": 0,
  "content": "<p>If it is not a secret, of course) Can we assume that public LB has great imbalance of <code>train_soundscapes</code> sites and private one has great imbalance of NON <code>train_soundscapes</code> sites ?</p>",
  "messages": [
    {
      "id": "1332162",
      "postDate": "06/02/2021 00:26:27",
      "content": "<p>If it is not a secret, of course) Can we assume that public LB has great imbalance of <code>train_soundscapes</code> sites and private one has great imbalance of NON <code>train_soundscapes</code> sites ?</p>",
      "rawMarkdown": "If it is not a secret, of course) Can we assume that public LB has great imbalance of `train_soundscapes` sites and private one has great imbalance of NON `train_soundscapes` sites ?",
      "votes": null
    },
    {
      "id": "1332167",
      "postDate": "06/02/2021 00:29:09",
      "content": "<p>When i was studying that file, i found that there are many bird species that had less than 5 occurences there, maybe their ratios were more extensive on the private lb?</p>",
      "rawMarkdown": "When i was studying that file, i found that there are many bird species that had less than 5 occurences there, maybe their ratios were more extensive on the private lb?",
      "votes": null
    },
    {
      "id": "1332170",
      "postDate": "06/02/2021 00:31:20",
      "content": "<p>Compare 28933_SSW and 31928_COR, for example. That giant of a variance between individual soundscapes should be a red flag, when you had only 80 soundscapes in the final test. For the subset of ~25 soundscapes in public test it is even worse.</p>\n<p>If the data was split random (as it should be!) - such big jumps are to be expected.</p>",
      "rawMarkdown": "Compare 28933_SSW and 31928_COR, for example. That giant of a variance between individual soundscapes should be a red flag, when you had only 80 soundscapes in the final test. For the subset of ~25 soundscapes in public test it is even worse.\n\nIf the data was split random (as it should be!) - such big jumps are to be expected.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1332167,
      "author_name": "victorasso",
      "author_url": "",
      "post_date": "06/02/2021 00:29:09",
      "content": "<p>When i was studying that file, i found that there are many bird species that had less than 5 occurences there, maybe their ratios were more extensive on the private lb?</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 1332170,
      "author_name": "fffrrt",
      "author_url": "",
      "post_date": "06/02/2021 00:31:20",
      "content": "<p>Compare 28933_SSW and 31928_COR, for example. That giant of a variance between individual soundscapes should be a red flag, when you had only 80 soundscapes in the final test. For the subset of ~25 soundscapes in public test it is even worse.</p>\n<p>If the data was split random (as it should be!) - such big jumps are to be expected.</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "1332162": "If it is not a secret, of course) Can we assume that public LB has great imbalance of `train_soundscapes` sites and private one has great imbalance of NON `train_soundscapes` sites ?",
    "1332167": "When i was studying that file, i found that there are many bird species that had less than 5 occurences there, maybe their ratios were more extensive on the private lb?",
    "1332170": "Compare 28933_SSW and 31928_COR, for example. That giant of a variance between individual soundscapes should be a red flag, when you had only 80 soundscapes in the final test. For the subset of ~25 soundscapes in public test it is even worse.\n\nIf the data was split random (as it should be!) - such big jumps are to be expected."
  },
  "source": "meta"
}